The objective of this proposal is to determine the frequency and types of random mutations in normal and malignant human cells. Our hypothesis is that normal mutation rates are insufficient to account for the multiple mutations observed in cancers and that cancer cells exhibit a mutator phenotype. Normal human cells are able to copy their genomes with only one, or at most a few, mistakes during each division cycle. In contrast, cancer cells exhibit large numbers of chromosomal aberrations and we postulate that they may harbor hundreds to thousands of additional, random changes in DNA sequence. Large numbers of random mutations (i.e., mutations that occur in one or only a few cells of a tumor) could contribute to tumor progression, and could account for the heterogeneity of cancer cells within a tumor, the rapid emergence of resistance to chemotherapy, and the ability of cancer cells to invade adjacent tissues and to metastasize. Quantitation of random mutation frequency in tumors may be useful in stratifying cancers, providing an index of tumor heterogeneity, and prognosticating malignant and metastatic potential and treatment outcome. We have established an assay that can detect one nucleotide substitution when present in one hundred million correct nucleotides. Our protocol allows us to measure the frequency of random changes in nucleotide sequence in introns and exons, and in dividing and non-dividing cells. We will use our assays to examine nucleotide changes in nuclear DNA from different normal human tissues, in different human cancers, and in human cells in culture. We will measure changes in mitochondrial DNA in prostatic cancers in relationship to tumor grade. If increases in mutation frequency drive tumor progression, it is extremely important to identify agents that can inhibit mutagenesis and thereby prevent or impede the progression of human cancers.